1 #ifndef MSP_STRINGCODEC_CODEC_H_
2 #define MSP_STRINGCODEC_CODEC_H_
9 namespace StringCodec {
21 Base class for string codecs. Use one of the derived classes or the function
22 create_codec to create a specific codec.
24 Unicode strings are represented as ustrings. An std::string is considered to
25 be an encoded sequence of bytes. A codec is able to determine if an encoded
26 string could be decoded with it.
32 Base class for string encoder.
34 Each codec class should contain an Encoder class derived from this. The
35 encode_char and transliterate functions must be overloaded. Some encoders
36 may find it useful or necessary to implement some other functions too
37 (particularly sync and reset for stateful codecs).
44 Encoder(ErrorMode em): err_mode(em==DEFAULT ? THROW_ON_ERROR : em) { }
46 virtual ~Encoder() { }
48 /** Encodes a single unicode character. If the character can't be
49 represented in this encoding, error() should be called. */
50 virtual void encode_char(unichar ch, std::string &buf) = 0;
52 /** Encodes a unicode string. This is equivalent to calling encode_char
53 for each character in the string with the same buffer. */
54 virtual void encode(const ustring &str, std::string &buf);
56 std::string encode(const ustring &);
58 /** Procuces a sequence of bytes that will bring the encoder back to the
60 virtual void sync(std::string &buf) { (void)buf; }
62 /** Resets the encoder to the initial state without producing output. */
63 virtual void reset() { }
66 /** Handles an error depending on the error mode.
68 THROW_ON_ERROR: throws err
69 IGNORE_ERRORS: does nothing
70 TRANSLITERATE: calls transliterate(ch, buf) */
72 void error(unichar ch, std::string &buf, const E &err)
74 if(err_mode==TRANSLITERATE)
75 transliterate(ch, buf);
76 else if(err_mode!=IGNORE_ERRORS)
80 /** Attempts to produce an alternative encoding for a unicode character.
81 Typically this includes dropping accent marks or romanizing letters. */
82 virtual void transliterate(unichar ch, std::string &buf) = 0;
86 Base class for string decoder.
88 Each codec class should contain an Decoder class derived from this.
95 Decoder(ErrorMode em): err_mode(em==DEFAULT ? THROW_ON_ERROR : em) { }
97 virtual ~Decoder() { }
99 /** Decodes a single character from a string. The iterator is advanced
100 to the next character. For stateful codecs, -1 may be returned if a
101 state change sequence was decoded but no character followed it. If
102 invalid input is encountered, the error() function should be called and
103 the iterator advanced only if it doesn't throw. */
104 virtual unichar decode_char(const std::string &str, std::string::const_iterator &i) = 0;
106 /** Decodes a string. */
107 virtual void decode(const std::string &str, ustring &buf);
109 ustring decode(const std::string &);
111 /** Resets the decoder to the initial state. */
112 virtual void reset() { }
115 /** Handles an error depending on the error mode. The return value is
116 suitable for returning from decode_char.
118 THROW_ON_ERROR: throws err
119 IGNORE_ERRORS: returns -1
120 TRANSLITERATE: return 0xFFFD */
122 unichar error(const E &err)
124 if(err_mode==TRANSLITERATE)
126 else if(err_mode==IGNORE_ERRORS)
138 /** Returns the name of the encoding handled by this codec. */
139 virtual const char *get_name() const = 0;
141 /** Creates an encoder for this codec. */
142 virtual Encoder *create_encoder(ErrorMode err_mode = DEFAULT) const = 0;
144 /** Creates a decoder for this codec. */
145 virtual Decoder *create_decoder(ErrorMode err_mode = DEFAULT) const = 0;
147 /** Determines whether the given string can be successfully decoded with
148 this codec. Note that this function returning true does not guarantee that
149 the string was actually encoded with this codec. In particular, many 8-bit
150 encodings are indistinguishable. */
151 virtual bool detect(const std::string &) const;
154 typedef Codec::Encoder Encoder;
155 typedef Codec::Decoder Decoder;
159 A helper class to provide some common functionality.
162 class StandardCodec: public Codec
168 StandardCodec(ErrorMode em): err_mode(em==DEFAULT ? THROW_ON_ERROR : em) { }
171 virtual Encoder *create_encoder(ErrorMode em = DEFAULT) const
172 { return new typename C::Encoder(em==DEFAULT ? err_mode : em); }
174 virtual Decoder *create_decoder(ErrorMode em = DEFAULT) const
175 { return new typename C::Decoder(em==DEFAULT ? err_mode : em); }
179 /** Convenience function that decodes a string. */
181 ustring decode(const std::string &s)
183 typename C::Decoder dec;
184 return dec.decode(s);
187 /** Convenience function that encodes a string. */
189 std::string encode(const ustring &s)
191 typename C::Encoder enc;
192 return enc.encode(s);
195 /** Convenience function that transcodes a string from one codec to another. */
196 template<typename F, typename T>
197 std::string transcode(const std::string &s)
199 return encode<T>(decode<F>(s));
202 /** Creates a codec for an encoding by name. The caller is responsible for
203 deleting the codec when it's no longer needed. */
204 Codec *create_codec(const std::string &);
206 /** Automatically detects the encoding of a string and creates a codec for it.
207 The codec must be deleted when it's no longer needed. */
208 Codec *detect_codec(const std::string &);
210 } // namespace StringCodec